All chronic liver diseases (causes: viral, alcohol, metabolic steatosis . . . ) are characterized by the development of liver lesions and the development of liver fibrosis. This liver fibrosis induces liver architecture modifications which are the source of major complications responsible for increased mortality. Fibrosis is progressive and usually reversible.
Quantification of hepatic fibrosis is important to determine the severity of liver disease, its prognosis and treatment indication. The quantification reference of hepatic fibrosis is a microscopical examination by an expert liver pathologist of a liver specimen usually obtained by liver biopsy (LB). Liver fibrosis is staged according to LB in semi quantitative fibrosis score.
Metavir classification is one of the most used classifications. It classifies liver fibrosis in five stages from F0 to F4, F4 stage corresponding to the final stage of cirrhosis. Fibrosis is interpreted as clinically significant when stage F≧2.
Currently the conventional reading of LB has limits. Indeed, these semi-quantitative scoring systems of fibrosis are limited by poor inter- or intra-observer reproducibility between pathologists (Rousselet et al, Hepatology, 2005, 41(2):257-264). The most important issue regarding the classification system by pathologists is that it is subjective. The limit of the poor reproducibility can be partially circumvented with the LB reading by an expert. However, a recent study showed that single experts did not reach excellent reproducibility whereas an expert panel seems to be the best reading reference (Boursier et al, BMC Gastroenterol, 2011, 11:132).
Thus, the challenging issue in pathology is to work with expert centers, with the probable risk of saturation. In addition, there is a rarefaction or increasing unavailability of pathologists.
There is thus a need for an image analysis method which is automated and thus allows the effective assessment of the presence and/or the severity of a lesion in an organ, such as, for example, the presence of liver fibrosis, or the prognosis (score indicative of an increased risk of mortality or of liver mortality or of hepatic complications).
Medical images analysis methods are known in the prior art.
For example, the US patent application US 2012/226709 describes an automated method for identifying prostate tissue samples in a database that are closest to a test prostate sample, in order to aid pathologists in diagnosing prostate cancer.
Moreover, the International patent application WO 2006/020627 describes a method for automated diagnosis of a disease, specifically of cancer, based on morphometric data extracted from tissue images by a computer. Examples of morphometric data include fractal dimension data, wavelet data and color channel histogram data.
In the field of hepatology, several methods of fibrosis morphometric measurements to describe its quantitative characteristics have been described:among different patterns, the area of fibrosis is the main characteristic. Fractal dimension of fibrosis, the perimeter and the size of collagenous elements were also described. For example, the European patent EP 2 021 933 describes a computerized in vitro diagnostic method for liver diseases, comprising measuring in a liver biopsy image, inflammation parameters (such as, for example, area of inflammatory tissue or the percentage of biopsy sample surface occupied by the inflammatory tissue) or fibrosis parameters (such as, for example, fractal dimension or corrected area of the fibrotic tissue).
However, these published techniques give quantitative information but are not well correlated with the diagnostic reference system of pathologists, i.e. for example in the field of hepatology, the METAVIR fibrosis score with 5 classes (F0: no fibrosis; F1: portal fibrosis without septa; F2: portal fibrosis with rare septa; F3: numerous septa without cirrhosis; F4: cirrhosis) and the Ishak staging system with 7 classes based on portal and septal fibrosis and the degree of completeness for cirrhosis.
At the difference of the current image analysis methods of the prior art, the Inventors herein developed a diagnostic method able to determine the fibrosis course that is well correlated with the expert's classification. So, they developed an automated measurement of many new morphometric patterns which describe quantitatively the information needed by pathologists when they make the METAVIR fibrosis (F) staging. All these measures may thus be used as a complement to classical optical diagnosis as required by the local conditions of clinical practice. Moreover, the method of the invention comprises the mathematical combination of these data in a score, thus allowing determining the Metavir stage with an automated measure and calculation or allowing diagnosing clinically significant fibrosis (CSF, METAVIR F≧2) and cirrhosis (METAVIR F4), with excellent accuracy and reproducibility.